By the way, I was too blasé in the grandparent comment. I have a model of TDT that does almost the right thing, but I haven’t figured out how to quine it so that it does exactly the right thing. (Technically, I can’t be sure it’ll still work if I do get the right quine in, but my gut feeling is that it will.)
You’ll see by Part III. (Hint: when you see an infinite regress, consider a clever quining.)
Ok. I see how to do quining to only cooperate with copies of your source code, but I don’t yet see how to do quining with outcome dependencies.
By the way, I was too blasé in the grandparent comment. I have a model of TDT that does almost the right thing, but I haven’t figured out how to quine it so that it does exactly the right thing. (Technically, I can’t be sure it’ll still work if I do get the right quine in, but my gut feeling is that it will.)
So, I’m going to be pretty disappointed if this whole affair is just someone inventing a meta-cliquebot that’s actually just a cliquebot.
Trust me, there’s better stuff than that. (In particular, I’ve got a more nicely formalized write-up of UDT. It’s just TDT I’m having issues with.)
The quining part usually seems to be the tricky part doesn’t it?